桥梁结构水下基础表观缺陷立体视觉检测试验

针对桥梁结构水下基础表观缺陷检测实际需求,提出立体视觉表观缺陷检测方法。采用SIFT算法完成表观缺陷图像特征提取与匹配,选用MATLAB编译立体视觉表观缺陷检测评定软件,集成桥梁结构水下基础表观缺陷立体视觉检测系统。使用该系统在实验室环境完成水及空气两种环境下混凝土剥落、裂缝及楔槽三种缺陷的检测对比试验,得到每种混凝土试块表观缺陷在不同环境下的检测结果。试验结果表明,立体视觉表观缺陷检测方法能准确评定表观缺陷几何尺寸,从而验证了该方法用于桥梁结构水下基础表观缺陷评定的实用性及可行性。

A Stereovision Detection Experiment for Apparent Defects in Underwater Foundation of Bridge Structure

To meet the practical detection requirements of apparent defect in underwater foundation of bridge structure, a stereovision-based apparent defect inspection method is proposed in this paper. First, the apparent defect image feature extraction and matching were automatically accomplished by adopting SIFT algorithm. Second, MATLAB was selected to compile the detection and evaluation software for stereovision apparent defect. Finally, a stereovision-based apparent defect detection and evaluation system was integrated for underwater foundation of bridge structure. By using this system, laboratory comparative experiments for detecting 3 defects including spalling, crack and wedge groove were carried out in water and air environment, respectively. Detection results of concrete specimen's every apparent defect in different environment were obtained. Results show that this stereovision-based apparent defect detection approach can accurately evaluate apparent defect geometry size. So the practicability and feasibility of this detection method is verified.